At Juspay, I architected a real-time Rust ingestion pipeline that streams Kafka events into Cassandra, ClickHouse, PostgreSQL, and AWS S3. I cut end-to-end ingestion latency by 90% and ClickHouse query time by 60%.
I also owned an analytics platform processing 40M+ daily events into partner-facing dashboards. I led a zero-downtime schema migration and built data revocation and record-merging logic with a full audit trail.
As a Software Engineer Intern at Amazon, I engineered an outage dashboard adopted by 250+ engineers and launched on Prime Day across 12 marketplaces. I consolidated 300+ levers into a MongoDB-backed source of truth, improving on-call efficiency by 40%.
On my AI On-Call Engineer project, I built a Python agent to triage Prometheus alerts and safely investigate Kubernetes incidents. In a benchmark across 50 faults, the best of three frontier LLMs resolved 97% of incidents with zero harmful actions.
